• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
11.[28 38]distribution of risk and return a statistical test of normality on dhaka stock exchange
 

11.[28 38]distribution of risk and return a statistical test of normality on dhaka stock exchange

on

  • 592 views

IISTE-call for paper www.iiste.org/Journals

IISTE-call for paper www.iiste.org/Journals

Statistics

Views

Total Views
592
Views on SlideShare
592
Embed Views
0

Actions

Likes
0
Downloads
11
Comments
0

0 Embeds 0

No embeds

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    11.[28 38]distribution of risk and return a statistical test of normality on dhaka stock exchange 11.[28 38]distribution of risk and return a statistical test of normality on dhaka stock exchange Document Transcript

    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012 Distribution of Risk and Return: A Statistical Test of Normality on Dhaka Stock Exchange Md. Mahbubul Haque Khan1* Umma Rumana Huq2 1. Department of Business Administration, East West University, Dhaka, Bangladesh 2. School of Business & Economics, United International University, Dhaka, Bangladesh * E-mail of the corresponding author: minmhk@gmail.comAbstractThe present study deals with the normal distribution of risk and return of the capital market of Bangladesh.Normal distribution of return is an essential assumption in the field of efficient market hypothesis whichposits that the returns of a market must follow the random walk behaviour. Again it is an integral part ofCapital Asset Pricing Model, which suggests that if an investor wants to get higher return then he mustconsider higher risk, this denotes to the normal distribution of risk and return. This study has used a set ofparametric and non-parametric tools to examine the returns calculated from the three indices of DhakaStock Exchange: DGEN (from 2002 to 2010), DSE20 (from 2002 to 2010) and DSI (from 2006 to 2010).As positive skewness and kurtosis are evident in most of the cases, the returns are found to be sufferingfrom some extremities. Daily, weekly and monthly returns are not normally distributed which shows thecontra-evidence of random walk behaviour of market return. Besides the inconsistency between risk andreturn (daily and weekly) is found, which suggests that additional return may be achieved without havingexposure to additional risk.Keywords: Dhaka Stock Exchange, Efficient Market Hypothesis, Kurtosis, Non-parametric Test, NormalDistribution, Parametric Test, Risk and Return, Skewness.1. IntroductionTest of normality is an important study in the field of statistical research, as there are so many statisticaltests and models which assume the sample data set to be normally distributed. There are some capitalmarket models which also assume that the data set (returns) must be normally distributed e.g., Capital AssetPricing Model (CAPM). This paper is going to focus on the normality assumption of Efficient MarketHypothesis (EMH), where EMH posits that stock market return must follow random walk behaviour, whichdenotes to the normal distribution of risk and return. Basically Efficient Market Hypothesis tells us that thesecurity prices fully reflect all available information (Fama, 1970). For this reason a person having a pieceof information cannot expect to predict the future price and have some abnormal gain on that information ifthe market is held to be efficient. So, the future price must be always random and unpredictable. A personcan only earn higher return by bearing higher risk. Hence the efficient market hypothesis can play animportant role in investment strategy formulation. In an inefficient market as the return distribution issomewhat skewed, there, the probability for a particular return can be assumed. The technicians by usingtheir technical trading rules can earn riskless profit in this circumstance. Investors can also be able to findout the implication of time horizon in their investment strategy to earn abnormal return. But in an efficientmarket condition, an investor has to bear an additional amount of risk for an additional amount of return.Basically here return is considered as a compensation for risk. The capital market of Bangladesh is anemerging market in South Asia. Day by day capital market is becoming an integral part of the economy ofBangladesh. Studies performed on market efficiency of the capital market of Bangladesh are not so many.And most of the articles present that the market is still not found weak form efficient (Mollah et al., 2005;Rahman and Hossain, 2006; Mobarek et al., 2008; Mollik and Bepari, 2009; and Nisar and Hanif, 2011).Besides some calendar anomalies like ‘Saturday Effect’, ‘April Effect’ and seasonality in monthly return 28
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012series (Hossain, 2004; and Bepari and Mollik, 2009) are also evident in the capital market of Bangladesh.As the evidence suggests that the market is not efficient in Bangladesh, the risk and return distribution willnot be normal. So basically this paper is going to examine whether and to what extent the risk and returnsare normally distributed in the capital market of Bangladesh within the sample time period. Again thispaper can also be able to validate some of the previous findings from the literature.2. Literature ReviewTesting normality and market efficiency of capital markets has been a prime area of interest for researchersfor a long time. In this part we are going to present some findings from the literature of both national(Bangladesh) and international (rest of the world) context. And at the end we will try to show the gap in theliterature and the justification for the present study.Aggarwal and Rivoli (1989) examined the efficiency in functioning of four large capital markets of Asia bystudying to find out seasonal and daily patterns in stock returns of these markets. With twelve years data(September 1976-June 1988), the study concluded on the existence of such seasonal patterns in thosemarkets. The study found out the existence of ‘January effect’ in markets of Hong Kong, Singapore andMalaysia. In addition to that, ‘day-of-the-week effect’ was also present in those markets in the form of lowMonday returns and strong ‘Tuesday effect’. Aparacio and Estrada (1997) strongly rejected the notion ofnormality in daily stock returns in their study on thirteen European securities market. The study used dataof period 1990-1995. By using time series of daily stock returns, they argued that normality can be assumedfor monthly returns but not for daily returns. Galai and Kedar-Levy (2002), in their working paper series,found the existence of day-of-the-effect on their study on Tel-Aviv stock exchange. They applied non-parametric tests and multiple comparison procedures on a data set of the period 1988-1998. The findingsrevealed that the first and last trading days’ returns were closely related than to any other day of the week.Aly, Mehdian and Perry (2004) analyzed Day-of-the-week effect on Egyptian stock market with a four-daytrading week. The study reported that returns do not vary among days in week indicating the absence of anysuch patterns. They also concluded on the weak form efficiency on Egyptian capital market. Naqvi (2004)tested the random-walk behavior of Karachi Stock Exchange in his study using KSE-100 index. Theresearch employed normality test, autocorrelation using Q-statistic and Dickey-Fuller test on the data set ofthe period 1994-2004. The Karachi Stock Exchange was found to be weak form efficient in both weeklyand monthly return series. Agrawal (2005) tested how the sample size can have an impact on the normalityassumption of stock returns and came up with the conclusion that it really had. According to his researchcarried out on the stock returns of Nifty and Sensex, daily stock returns with large sample size were notnormally distributed while monthly stock returns with small sample size were normally distributed. Onderand Zaman (2005) examined how the test statistics for the normality test of errors can have a differentresult if the regression residuals from robust regression are used in place of OLS residuals. They actuallyfound out that such use leads to significantly superior ability in detecting lack of normality in suitablesituations. O. (2008) attempted to test normality and market efficiency on Nigerian Stock Exchange. Heexamined that though in recent years Nigerian stock exchange has been growing rapidly in terms of volumetrading and market capitalization; the market was not weak form efficient the stock returns were notnormally distributed. The study was conducted using monthly all share index from January 1993 toDecember 2007 and the methods used for the analysis included Skewness, Kurtosis, Jarque-Bera, Runs testetc. He added to recommend reducing transaction costs and minimizing institutional restrictions on tradingto improve market activities. Das and Bhattacharjee (2009) attempted to build up a graphical tool to checkand visualize the normality of a dataset using empirical distribution function and K-S test statistic. Jeyanthi(2010) tested the weak-form efficiency of Asian emerging stock markets. The study used daily priceindices of China, Indonesia, Kuala Lumpur, Korea, Taiwan and India for the period of April 1998 to March2009. The study showed varying results as the K-S test, runs test, autocorrelation test and LB test provedthat these markets were not weak-form efficient but on the other hand unit root test significantly proved theexistence of weak-form efficiency among the markets. Kumar and Dhankar (2011) applied a set ofparametric and non-parametric tests on Indian Stock Market to find out whether the risk and return of thatmarket are distributed normally. They examined three listed indices of Bombay Stock Exchange for theperiod 1996 to 2006 and found the evidence of non-normality in case of daily and weekly returns but 29
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012normality in case of monthly and annual returns. Besides their study pointed out negative skewness inreturns and emphasized on time horizon in investment strategy.Such research on testing of normality and market efficiency has also been practiced on the stock market ofBangladesh namely Dhaka Stock Exchange (DSE). Hossain (2004) studied on the small banking portfoliosin Bangladesh to find out any existence of days-of-the-week effect. The study revealed the existence of‘Saturday Effect’ as Saturday was the then first trading day following the weekend. The study alsoidentified the existence of ‘beginning of the week effect’ while buying and ‘weekend effect’ while sellingsecurities. Rahman, Salat and Bhuiyan (2004) proved in their study that Dhaka Stock Exchange is weak-form efficient. They used monthly index of DSE for the period of 1990-2003 to apply unit root test andcame up to support the null hypothesis of random walk model. Mollah, Rahman and Islam (2005) analyzedthe behavior of daily DSE 20 index and prices of the securities. The ARMA, ARIMA, ACF, PACF andDimson’s Market Model identified that the return series was a stationary time series and the market was notweak-form efficient. Islam and Khaled (2005) applied heteroscedasticity-adjusted methods to test for weak-form efficiency of Dhaka Stock Exchange as historic study results gave conflicting evidences on weak formefficiency. They found that if the heteroscedasticity-robust test is used instead of an autocorrelation testthen the result can be transformed from weak form inefficiency to weak form efficiency. Rahman andHossain (2006) carried out a study to find whether Dhaka Stock Exchange is weak-form efficient assumingthe data set is normally distributed. They applied both parametric and non-parametric tests on tworenowned indices of the market from 1994-2005. The study concluded on the ground that the daily returnseries was not normally distributed and the market was not weak-form efficient. Mobarek et al. (2008)sought to find evidence on the randomness and independence of the return series of Dhaka Stock Exchange.Various statistical models like K-S test, Runs test, Auto-Correlation test, ARIMA model etc were run onthe daily price index from 1988 to 2000. The study found that security returns did not follow the randomwalk model. The existence of significant auto-correlation coefficient at different time lags proved themarket to be weak-form inefficient. She also added that such inefficiency might be due to delay indissemination of price-sensitive information or biases in the responses of market participants. Bepari andMollik (2009) investigated the presence of seasonality in monthly return series of DSE All Share PriceIndex for the period from 1993 to 2006. Their research confirmed the existence of monthly effect in theform of ‘April effect’ instead of ‘January effect’. They also added to confirm that the market was not weak-form efficient meaning investors could earn above average income. Rahman, Ara and Zheng Sr. (2009)used a variety of volatility models such as GARCH model, Random walk model and Autoregressive modelto test the volatility of Dhaka Stock Exchange returns from 1999 to 2004. The study found the marketreturn to be more volatile after 27 November, 2001. Mollik and Bepari (2009) examined the weak-formefficiency of Dhaka Stock exchange by adjusting for the thin trading problem. The data set covered DSEGeneral Index and DSE 20 index for the period of 2002-2007. The study found that the returns are notnormally distributed and the market failed to be weak-form efficient. Nisar and Hanif (2011) tested theweak-form efficiency on four emerging stock markets in Asia including Bangladesh. According to theirRuns test, Durbin-Watson test and Unit Root test which they applied on the monthly, weekly and dailyreturn series of the markets from 1997 to 2011, none of the markets showed to be weak form efficient.Mollik and Bepari (2011) attempted to measure the risk-return relationship in Dhaka Stock exchange. Theirstudy found pretty straight, anticipated positive relationship between risk and return for both single securityand portfolio. Still the less significant positive risk-return relationship for high risk portfolios might suggestexistence of mispricing or market anomalies.From the literature it is quite apparent that on the capital market of Bangladesh there is a lack of studyparticularly on risk and return distribution or normality test, though we know that this is a pre-requisite forso many statistical and financial models. Some researchers have also worked on normality test as a part oftheir research. Again for testing the normality of return basically researchers have commonly chosenKolmogorov Smirnov test as a non-parametric test (Rahman and Hossain, 2006; Mobarek et al., 2008; andKumar and Dhankar, 2011). But Razali and Wah (2011) compared among the Shapiro-Wilk test,Kolmogorov-Smirnov test, Lilliefors and Anderson-Darling test and found that irrespective of thedistribution of observations Shapiro Wilk test is the most powerful test and Kolmogorov Smirnov is the 30
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012least powerful test, besides Lilliefors test always outperforms the Kolmogorov Smirnov test. So the presentstudy has incorporated Kolmogorov-Smirnov test with Lilliefors correction and Shapiro Wilk test forconducting the research. We have also found that Kumar and Dhankar (2011) have used F-test or Analysisof Variance test to find out the distribution of risk. But F-test assumes the sample observations to benormally distributed, so in this research we have used Kruskal-Wallis test (a non-parametric test which canbe compared with one way ANOVA test) along with F-test. Moreover the present study has completelyconcentrated on test of normality of risk and return of the capital market of Bangladesh.3. Methodology and Data SourcesReturns calculated from the capital market indices are widely used as the proxies for the market return.Dhaka Stock Exchange (DSE) has three indices to represent the capital market of Bangladesh. In this studywe have used the three indices of DSE to calculate the market return on a daily, weekly, monthly andannual basis. DSE General Index (DGEN) consists of A, B, G and N categories of listed shares of DhakaStock Exchange. In our study we have considered DGEN from January 2002 to December 2010. DSE20Index consists of 20 blue chip shares listed in DSE. The sample period for DSE20 is January 2002-December 2010. DSE All Share Price Index (DSI) includes all the shares listed in DSE. The authority ofDSE stopped calculating this index from 08 December 2003. The authority reintroduced this index on 28March 2005. So there was a discontinuity in DSI index and finally here we have considered DSI fromJanuary 2006-December 2010 as the sample period.Besides, we have also considered three sub-periods (2002-2004, 2005-2007, and 2008-2010) of DGEN andDSE20 to find out whether the samples have emerged from the same population. During the whole period2002-2010 we see (Graph 1) the capital market of Bangladesh had an increasing trend. From 2002 to 2004the market was quite flat and during 2005-2007 the market experienced some fluctuations. But during theperiod 2008 to 2010 the market really observed a tremendous growth. Basically that time the market washighly influenced by the IPO of Grameen Phone Ltd. (the largest mobile phone operator of Bangladesh).Daily market return jumped at 20.38% on 16 November 2009 after the debut of Grameen Phone share onDhaka Stock Exchange, which also pulled the market index to a record high. The Financial Express, apopular financial newspaper of Bangladesh reported on 17 November 2009 in the following way:‘Shares of Grameen Phone (GP) climbed 2.5 times higher than its face value at the Dhaka Stock Exchange(DSE) Monday after the countrys largest mobile phone operator made debut, sending the indices to arecord high’.In this study we have used natural logarithmic return for the ease and simplicity of use and calculated in thefollowing way:Return on index at time period t = ln ( It / It-1 ), where It is the index price at time period t and It-1 is the priceof index at time period previous t. We have considered only the changes in stock price and overlooked anydividend payment.We have used several parametric and non-parametric tests to determine whether the risk and returncalculated from our samples are normally distributed. We have used Kolmogorov Smirnov (K-S test withLilliefors Significance Correction) test and Shapiro Wilk test to find out whether the returns calculatedfrom our sample data fit normal distribution. The significance of the test statistic will reject the assumptionof normality in respective sample data (Pallant, 2005). One way ANOVA (Analysis of Variance or F-test)and Kruskal-Wallis (K-W) tests are applied on returns calculated from the three sub-periods’ sample data tofind out whether these have emerged from the same population. If these are found not to be emerged fromthe same population then we have concluded that risk is not normally distributed (Kumar and Dhankar,2011). All the tests are assumed at 5% level of significance.4. Major Findings4.1 Risk and Return Distribution of Daily Return (2002-2010)In this part basically we have three statistical tables presenting the facts related to normality of daily returnof Dhaka Stock Exchange. If we have a look on Table 1 then we find that the index DGEN has a positiveskewness whereas DSE20 index is negatively skewed. Kurtosis for both the indices is greater than zero. So 31
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012from the perspective of skewness and kurtosis we find a violation of normality. Again test statistics for K-Sand Shapiro Wilk tests are significant which means the return distribution is not normal for the daily returnwithin the sample period. This finding is also consistent with other findings where the day of the weekeffect on this market is evident.From Table 2 we get the statistical summary of daily returns from 2002 to 2010 for three non-overlappingsub-periods- 2002-2004, 2005-2007 and 2008-2010. DGEN exhibits negative skewness in 2005-2007period and positive skewness for the other two periods. Daily return from DSE20 index is negativelyskewed for all the periods. Kurtosis is higher than zero for all the sub-periods, which is a symptom of non-normality. If there is higher kurtosis than normal, then more of the variability can be explained by the fewerobservations’ extreme differences from the mean. And we observe kurtosis for DGEN in 2008-2010 isabnormally high. The same is not observed in DSE20. So, the returns of the securities beyond DSE20 areheld liable for this abnormality. Information from the market suggests, this has happened for the IPO ofGrameen Phone Ltd. (the largest mobile phone operator in Bangladesh) within that sub-period. For bothDGEN and DSE20 indices K-S and Shapiro Wilk test statistics are significant. So these show non-normality in daily return of each sub-period.As we see that the Levene Statisitc in Table 3 is significant for both the indices, F-value will not be able toprovide a correct decision about the sample. So we need to rely on Welch and Brown-Forsythe test statisticfor the decision. And we find both the tests indicate that the three samples have emerged from the samepopulation, hence the risk for the three sub-periods will not significantly differ from each other. By beingnot significant at 5% level of significance, Kruskal-Wallis (K-W in Table 2) test statistic supports the samefinding.4.2 Risk and Return Distribution of Weekly Return (2002-2010)This part deals with the risk and return distribution of weekly return of Dhaka Stock Exchange. Table 4shows some statistics of weekly return of DGEN and DSE20 during the period 2002-2010. DGEN showshigher weekly return than that of DSE20. Skewness and kurtosis are positive for both the indices, thisdenotes to non-normality. Kurtosis in weekly return has been reduced than that we observed in daily return(40.512). So it is quite apparent that due to any abnormally high daily return that situation emerged. Butstill K-S and Shapiro Wilk tests results are significant which tells us that weekly returns are not normallydistributed.From Table 5 we see that the K-S and Shapiro Wilk tests are not significant for DGEN in the sub-period2005-2007 and for DSE20 in the sub-period 2008-2010, which denotes to normal distribution of returnduring the respective sub-period for the relevant index. As the Levene Statistic (Table 6) is not significantat 5% level, we can rely on the F-value. F-statistic and K-W value suggest that risk is normally distributedfor both the indices during the sample period.4.3 Risk and Return Distribution of Monthly Return (2002-2010)Table 7 indicates the statistical summary of Monthly Return for the sample period 2002-2010 for DGENand DSE20. Here the K-S and Shapiro Wilk Statistics are not significant for both of the indices. Thissignals that the monthly returns are not normally distributed. But if we look at the sub-periods then we findexcept for DSE20 in 2002-2004 sub-period, all the sub-periods of both the indices are normally distributed.From the F-value and K-W statistic (Table 8) we find the sub-periods for both the indices have emergedfrom the same population, because we see the test results are not significant at 5% level of significance,which also tells about the normal distribution of risk during the sample period. We can rely on the F-valueas Levene test statistic from Table 9 is not significant.4.4 Risk and Return Distribution of Annual Return (2002-2010)Table 10 presents the statistical summary of the annual returns of two indices for the sample period (2002-2010). DGEN index provides higher average return than the DSE20 index, though DGEN has a lowerstandard deviation. As the K-S and Shapiro Wilk test results are not significant at 5% level of significance,we can assume the annual returns of the capital market are normally distributed. But still as the number ofobservations is too low, we should not rely on these statistics. 32
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 20124.5 Risk and Return Distribution of DSI (2006-2010)Earlier in methodology part we mentioned that due to unavailability of data we could not include anotherstock market index DSI (All Share Price Index) of Dhaka Stock Exchange in our previous analysis. But inthis part we are presenting it separately for the time period of 2006-2010. Table 11 presents that theskewness of daily and weekly returns are positive but monthly return is negative. Kurtosis of daily return isabnormally high which also supports the findings from daily return of DGEN for the period of 2008-2010.Shapiro Wilk test statistics are significant for all the daily, weekly and monthly returns. So returns are notnormally distributed.5. ConclusionThis study has been conducted to examine the risk and return distribution of the stock market ofBangladesh. This paper has come up with some important points to be noted. In most of the cases returnsare found to be positively skewed, which denotes that most of the returns are concentrated on the left sideof the mean with some extreme values to the right. We have also noticed on 16 November 2009 marketreturn was 20.38% which is really abnormally high, and it happened due to the debut of Grameen PhoneLtd. shares to the market. The standard deviation overestimates the downside risk and shows a biased resultdue to this positive skewness. We have found kurtosis is too high for the indices DGEN and DSI whichagain tells us that more of the variability resulted from the extreme deviations of fewer observations fromthe mean. From the indices DGEN and DSI it has become evident that the daily, weekly and monthlyreturns are not normally distributed, which denies the random walk behaviour of stock price. This complieswith the previous findings of some researchers who have found calendar anomalies like ‘Saturday Effect’,‘April Effect’ and seasonality in monthly return series (Hossain, 2004; and Bepari and Mollik, 2009).Again we have observed risks have been distributed normally in all the cases but daily and weekly returnshave not been normally distributed, so here we find an asymmetry between risk and return relationship. Soinvestors may have the chance to earn abnormal return i.e., return without bearing sufficient risk. Hence themarket is not efficient which also supports some past studies (Mollah et al., 2005; Rahman and Hossain,2006; Mobarek et al., 2008; Mollik and Bepari, 2009; and Nisar and Hanif, 2011).ReferencesAggarwal, Reena, & Rivoli, Pietra. (1989). Seasonal and Day-of-the-Week Effects in Four Emerging StockMarkets. The Financial Review, 24(4), 541-550.Agrawal, Gaurav. (2005). Impact of Sample Size on the Distribution of Stock Returns- An Investigation ofNifty & Sensex. Paper Presented at 9th Capital Markets Conference of Indian Institute of Capital Markets.Aly, Hassan, Mehdian, Seyed, & Perry, Mark J. (2004). An Analysis of Day-of-theWeek Effects in theEgyptian Stock Market. International Journal of Business, 9(3), 301-308.Aparicio, Felipe Miguel, & Estrada, Javier. (1997, April). Empirical Distribution of Stock Returns:European Securities Market, 1990-95. (Working Paper 97-23). Madrid, Spain: Departmento de Economiade la Empresa. Available at SSRN: http://ssrn.com/abstract=40558.Bepari, Khokan, & Mollik, Abu Taher. (2009). Seasonalities in the Monthly Stock Returns: Evidence fromBangladesh Dhaka Stock Exchange (DSE). International Research Journal of Finance and Economics, 24,167-176.Das, Kishore K., & Bhattacharjee, Dibyojyoti. (2009). On Using Empirical Distribution Function Plot forChecking the Normality Assumption of a Data Set. Assam Statistical Review, 22(1), 50-58.Debutant GP pulls DGEN to all-time high level. (2009, November 17). The Financial Express, Bangladesh.Fama, Eugene F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. Journal ofFinance, 25(2), 383-417. 33
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012Galai, Dan, & Kedar-Levy, Haim. (2002, February). Multiple Comparisons of Return Distributions ofReturn Distributions: A New Look at the Day-of-the-Week Effect. Available at SSRN:http://ssrn.com/abstract=302959.Hossain, Farhad. (2004). Days of the Week Effect in Dhaka Stock Exchange: Evidence from SmallPortfolios of Banking Sector. Jahangirnagar Review, Part II: Social Science, XXVIII, 73-82.Islam, Ainul, & Khaled, Mohammed. (2005). Tests of Weak-Form Efficiency of the Dhaka StockExchange. Journal of Business Finance & Accounnting, 32(7-8), 1613-1624.Jeyanthi, B. J. Queensly. (2010). Weak-Form Market Efficiency in India and Its Emerging AsianCounterparts. IUP Journal of Public Finance, 8(4), 58-70.Kumar, Rakesh, & Dhankar, Raj S. (2011). Distribution of Risk and Return: A Test of Normality in IndianStock Market. South Asian Journal of Management, 18(1), 109-118.Mobarek, Asma, Mollah, A. Sabur, & Bhuyan, Rafiqul. (2008). Market Efficiency in Emerging StockMarket: Evidence from Bangladesh. Journal of Emerging Market Finance, 7(1), 17-41.Mollah, A. Sabur, Rahman, Muhammad Zahedur, & Islam, Mohammad Saiful. (2005). Return Behavior ofthe DSE-20: An Empirical Investigation on the Dhaka Stock Exchange. Journal of Business Studies, 26(2),151-162.Mollik, Abu Taher, & Bepari, M Khokan. (2009, August 24). Weak-Form Market Efficiency of DhakaStock Exchange (DSE), Bangladesh. Paper presented at the 22nd Australasian Finance and BankingConference 2009.Mollik, Abu Taher., & Bepari, Khokon. (2011, August). Risk-Return Trade-Off in Dhaka Stock Exchange,Bangladesh: An Emerging Market Evidence. Available at SSRN: http://ssrn.com/abstract=1917337.Naqvi, Syed Ali Abbas. (2004, September). Does KSE-100 Index Follows a Random Walk: An EmpiricalStudy. BIMCS Department of Management Sciences Working Paper. Available at SSRN:http://ssrn.com/abstract=693261.Nisar, Saqib, & Hanif, Muhammad. (2011, November). Testing Weak Form of Efficient MarketHypothesis: Empirical Evidence from South Asia. World Applied Science Journal, Forthcoming. Availableat SSRN: http://ssrn.com/abstract=1962210.O., Emenike Kalu. (2008, October). Empirical Test for Weak Form Efficient Market Hypothesis of theNigerian Stock Exchange. Working Paper Series. Available at SSRN: http://ssrn.com/abstract=1291273.Onder, A. Ozlem, & Zaman, Asad. (2005). Robust Tests for Normality of Errors in Regression Models.Economics Letters, 86, 63-68.Pallant, Julie. (2005). Descriptive Statistics. In SPSS Survival Manual: A Step by Step Guide to DataAnalysis Using SPSS (2nd ed., pp. 53-58). Australia: Ligare, Sydney.Rahman, Mohammad Masudur, Ara, Laila Arjuman, & Zheng Sr., Zhenlong. (2009, April). Jump, Non-Normal Error Distribution and Stock Price Volatility- A Non-paramedic Specification Test. The SingaporeEconomic Review, Available at SSRN: http://ssrn.com/abstract=1500988.Rahman, Muhammad Zahedur, Salat, Amirus, & Bhuiyan, Mohammad Mominul Hoque. (2004). TestingWeak-Form Efficiency of the Dhaka Stock Exchange. Journal of Business Studies, 25(2), 175-188.Rahman, Shofiqur, & Hossain, Farhad. (2006). Weak-Form Efficiency: Testimony of Dhaka Stockexchange. Journal of Business Research, 8, 1-12.Razali, Nornadiah Mohd, & Wah, Yap Bee. (2011). Power Comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and Anderson-Darling Tests. Journal of Statistical Modeling and Analysis, 2(1), 21-33. 34
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012Table 1. Daily Return (2002-2010) Daily Return (2002-2010) Index Mean Standard Skewness Kurtosis K-Sa Shapiro Wilk Return % Deviation % DGEN .10035 1.1815 2.043 40.512 .072* .878* DSE20 .06906 1.1495 -0.144 4.934 .068* .945* Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 2. Statistical Summary (Daily Return) Statistical Summary (Daily Return) Index Period Mean Standard Skewness Kurtosis K-Sa Shapiro F K-W Return Deviation Wilk % % DGEN 2002- .1023 .83732 0.042 3.895 .068* .950* .843** 2.73** 2004 2005- .0589 1.2515 -0.193 4.081 .055* .960* 2007 2008- .1394 1.4255 3.784 56.631 .087* .801* 2010 DSE20 2002- .0833 .89790 -0.108 5.957 .077* .926* 1.03** 3.44** 2004 2005- .0195 1.2754 -0.211 4.394 .065* .951* 2007 2008- .1019 1.2703 -0.061 3.754 .053* .963* 2010 Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 3. Testing the Validity of ANOVA (Daily Return) Testing the Validity of ANOVA (Daily Return) Index Test of Homogeneity of Robust Tests of Equality of Means Variances Levene Statistic Welch Brown-Forsythe DGEN 37.755* .669** .811** DSE20 35.095* .879** 1.001** Note: * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 4. Weekly Return (2002-2010) Weekly Return (2002-2010) Index Mean Standard Skewness Kurtosis K-Sa Shapiro Wilk Return % Deviation % DGEN .48880 2.6132 1.052 7.323 .062* .941* DSE20 .33671 2.7103 0.482 1.660 .057* .975* Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significance 35
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012Table 5. Statistical Summary (Weekly Return) Statistical Summary (Weekly Return) Index Period Mean Standard Skewness Kurtosis K-Sa Shapiro F K-W Return Deviation Wilk % % DGEN 2002- .54876 2.2959 0.618 2.424 .109* .940* .874** 1.07** 2004 2005- .27000 2.5724 -0.101 1.163 .062** .985** 2007 2008- .64662 2.9323 1.998 11.958 .088* .883* 2010 DSE20 2002- .44811 2.5676 0.828 2.038 .092* .959* 1.032** 3.07** 2004 2005- .08267 2.7801 0.419 2.610 .077* .956* 2007 2008- .47845 2.7768 0.306 0.563 .069** .985** 2010 Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 6. Testing the Validity of ANOVA (Weekly Return) Testing the Validity of ANOVA (Weekly Return) Index Test of Homogeneity of Robust Tests of Equality of Means Variances Levene Statistic Welch Brown-Forsythe DGEN 2.348** 0.841** 0.874** DSE20 0.685** 0.999** 1.032** Note: * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 7. Monthly Return (2002-2010) Monthly Return (2002-2010) Index Mean Standard Skewness Kurtosis K-Sa Shapiro Wilk Return % Deviation % DGEN 2.2467 6.9290 0.182 1.743 .090* .974* DSE20 1.5971 7.1425 0.459 1.254 .107* .972* Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significance 36
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012Table 8. Statistical Summary (Monthly Return) Statistical Summary (Monthly Return) Index Period Mean Standard Skewness Kurtosis K-Sa Shapiro F K-W Return Deviation Wilk % % DGEN 2002- 2.7645 5.5627 0.621 -0.044 .145** .958** .636** .88** 2004 2005- 1.1823 7.5129 -0.594 1.664 .140** .949** 2007 2008- 2.8077 7.5594 0.878 1.639 .116** .951** 2010 DSE20 2002- 2.3679 6.1694 0.977 0.604 .174* .923* .786** 1.74** 2004 2005- 0.3920 7.9709 0.558 2.637 .146** .944** 2007 2008- 2.0528 7.1899 0.271 -0.010 .094** .987** 2010 Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 9. Testing the Validity of ANOVA (Monthly Return) Testing the Validity of ANOVA (Monthly Return) Index Test of Homogeneity of Robust Tests of Equality of Means Variances Levene Statistic Welch Brown-Forsythe DGEN 0.618** 0.592** 0.639** DSE20 0.333** 0.723** 0.457** Note: * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 10. Annual Return (2002-2010) Annual Return (2002-2010) Index Mean Standard Skewness Kurtosis K-Sa Shapiro Wilk Return % Deviation % DGEN 25.5111 34.7213 0.195 -2.110 .198** .867** DSE20 17.5554 35.0942 0.374 -1.411 .199** .912** Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significanceTable 11. DSI Return (2006-2010) DSI Return (2006-2010) Index Mean Standard Skewness Kurtosis K-Sa Shapiro Wilk Return % Deviation % Daily .14180 1.3102 2.506 38.414 .061* .875* Weekly .64306 2.7045 1.411 7.857 .053** .928* Monthly 2.0721 7.7413 -0.275 2.013 .094* .959* Note: a Lilliefors Significance Correction * Significant at 5% level of significance; ** Not significant at 5% level of significance 37
    • Research Journal of Finance and Accounting www.iiste.orgISSN 2222-1697 (Paper) ISSN 2222-2847 (Online)Vol 3, No 3, 2012Graph 1. DSE General Index (DGEN from January 2002-December 2010) 38
    • International Journals Call for PaperThe IISTE, a U.S. publisher, is currently hosting the academic journals listed below. The peer review process of the following journalsusually takes LESS THAN 14 business days and IISTE usually publishes a qualified article within 30 days. Authors shouldsend their full paper to the following email address. More information can be found in the IISTE website : www.iiste.orgBusiness, Economics, Finance and Management PAPER SUBMISSION EMAILEuropean Journal of Business and Management EJBM@iiste.orgResearch Journal of Finance and Accounting RJFA@iiste.orgJournal of Economics and Sustainable Development JESD@iiste.orgInformation and Knowledge Management IKM@iiste.orgDeveloping Country Studies DCS@iiste.orgIndustrial Engineering Letters IEL@iiste.orgPhysical Sciences, Mathematics and Chemistry PAPER SUBMISSION EMAILJournal of Natural Sciences Research JNSR@iiste.orgChemistry and Materials Research CMR@iiste.orgMathematical Theory and Modeling MTM@iiste.orgAdvances in Physics Theories and Applications APTA@iiste.orgChemical and Process Engineering Research CPER@iiste.orgEngineering, Technology and Systems PAPER SUBMISSION EMAILComputer Engineering and Intelligent Systems CEIS@iiste.orgInnovative Systems Design and Engineering ISDE@iiste.orgJournal of Energy Technologies and Policy JETP@iiste.orgInformation and Knowledge Management IKM@iiste.orgControl Theory and Informatics CTI@iiste.orgJournal of Information Engineering and Applications JIEA@iiste.orgIndustrial Engineering Letters IEL@iiste.orgNetwork and Complex Systems NCS@iiste.orgEnvironment, Civil, Materials Sciences PAPER SUBMISSION EMAILJournal of Environment and Earth Science JEES@iiste.orgCivil and Environmental Research CER@iiste.orgJournal of Natural Sciences Research JNSR@iiste.orgCivil and Environmental Research CER@iiste.orgLife Science, Food and Medical Sciences PAPER SUBMISSION EMAILJournal of Natural Sciences Research JNSR@iiste.orgJournal of Biology, Agriculture and Healthcare JBAH@iiste.orgFood Science and Quality Management FSQM@iiste.orgChemistry and Materials Research CMR@iiste.orgEducation, and other Social Sciences PAPER SUBMISSION EMAILJournal of Education and Practice JEP@iiste.orgJournal of Law, Policy and Globalization JLPG@iiste.org Global knowledge sharing:New Media and Mass Communication NMMC@iiste.org EBSCO, Index Copernicus, UlrichsJournal of Energy Technologies and Policy JETP@iiste.org Periodicals Directory, JournalTOCS, PKPHistorical Research Letter HRL@iiste.org Open Archives Harvester, Bielefeld Academic Search Engine, ElektronischePublic Policy and Administration Research PPAR@iiste.org Zeitschriftenbibliothek EZB, Open J-Gate,International Affairs and Global Strategy IAGS@iiste.org OCLC WorldCat, Universe Digtial Library ,Research on Humanities and Social Sciences RHSS@iiste.org NewJour, Google Scholar.Developing Country Studies DCS@iiste.org IISTE is member of CrossRef. All journalsArts and Design Studies ADS@iiste.org have high IC Impact Factor Values (ICV).